National Tax Journal Vol 49 no. 4 (December 1996) pp.

617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

TAXATION AND ECONOMIC GROWTH

ERIC ENGEN * & JONATHAN SKINNER
**

Abstract - Tax reforms are sometimes touted as having strong macroeconomic growth effects. Using three approaches, we consider the impact of a major tax reforma 5 percentage point cut in marginal tax rateson long-term growth rates. The first approach is to examine the historical record of the U.S. economy to evaluate whether tax cuts have been associated with economic growth. The second is to consider the evidence on taxation and growth for a large sample of countries. And finally, we use evidence from microlevel studies of labor supply, investment demand, and productivity growth. Our results suggest modest effects, on the order of 0.2 to 0.3 percentage point differences in growth rates in response to a major tax reform. Nevertheless, even such small effects can have a large cumulative impact on living standards.

tax rates, or by replacing the current federal income tax with a consumptiontype tax, the United States can experience increased work effort, saving, and investment, resulting in faster economic growth. For example, Steve Forbes vaulted briefly into the political limelight based almost solely on his advocacy of a flat tax which cut nearly every persons tax bill, but which was supposed to balance the budget by stimulating economic growth. The Kemp Commission suggested that its general principles for tax reform would almost double U.S. economic growth rates over the next five to ten years.1 Most recently, presidential candidate Robert Dole proposed a 15 percent across-the-board income tax cut coupled with a halving of the tax on capital gains, with a predicted increase in gross domestic product (GDP) growth rates from about 2.5 to 3.5 percentage points. Others have questioned whether tax reform would have such beneficial effects on economic growth.2 If tax cuts fail to produce the projected boost in economic growth, tax revenues could decline, putting upward pressure on the deficit, worsening levels of national saving, and leading to laggard economic growth in the future. At this stage, however, there is little agreement about
617

INTRODUCTION By now, a presidential campaign is incomplete without at least one proposal for tax reform. Recent proposals suggested that by reducing marginal
*

Federal Reserve Board, Washington, D.C. 20551. Department of Economics, Dartmouth College, Hanover, NH 03755, and NBER, Cambridge, MA 02138.
**

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

whether a major tax reform would provide an economic boon to the United States or impede economic growth. In this paper, we reexamine the relationship between economic growth and taxation in light of the accumulated economic evidence, both from the United States and other countries. While many economists would agree with the proposition that high taxes are bad for economic growth, we show that this proposition is not necessarily obvious, either in theory or in the data. However, we find that the evidence is consistent with lower taxes having modest positive effects on economic growth. While such growth effects are highly unlikely to allow tax cuts to pay for themselves, they can contribute to substantial differences in the level of economic activity and living standards, particularly over the long term. SHOULD WE EXPECT TAXES TO AFFECT GROWTH? A THEORETICAL PERSPECTIVE Before jumping into the morass of empirical evidence, it is useful to first ask the question: How does tax policy affect economic growth? By discouraging new investment and entrepreneurial incentives? By distorting investment decisions because the tax code makes some forms of investment more profitable than others? Or by discouraging work effort and workers acquisition of skills? These questions are often addressed in an accounting framework first developed by Solow (1956). In this approach, the output, y, of an economy, typically measured by GDP, is determined by its economic resourcesthe size and skill of its workforce, m, and the size and technological productivity of its capital stock, k. Thus, a country like the United States might be expected to have a greater per capita output than
618

Mozambique because its (per capita) capital stock is so much larger and more technologically advanced and its workers have more skills, or human capital. The growth rate of economic output therefore will depend on the growth rate of these resources physical capital and human capitalas well as changes in the underlying productivity of these general inputs in the economy. More formally, we can decompose the growth rate of the economys output into its different components:

1 . . . y i = ik i + im i + i where the real GDP growth rate in . country i is denoted yi and the net investment rate (expressed as a fraction of GDP), equivalently the change over . time in the capital stock, is given by ki . The percentage growth rate in the effective labor force over time is written . mi, while the variable i measures the economys overall productivity growth. There are two other relevant variables in equation 1, which are the coefficients measuring the marginal productivity of capital, i, and the output elasticity of labor, i.3 For example, if there were a one percentage point increase in the growth rate of the (skill-adjusted) labor force and were equal to 0.75, the implied increase in the economic growth rate would be 0.75 percentage point. Alternatively, if the investment rate were to rise by one percentage point and were 0.10, the growth rate of output would rise by 0.10 percentage point. This theoretical framework allows us to catalog the five ways that taxes might affect output growth, corresponding to each of the variables on the right-hand side of equation 1. First, higher taxes

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

can discourage the investment rate, or . the net growth in the capital stock (ki in equation 1 above), through high statutory tax rates on corporate and individual income, high effective capital gains tax rates, and low depreciation allowances. Second, taxes may attenu. ate labor supply growth mi by discouraging labor force participation or hours of work, or by distorting occupational choice or the acquisition of education, skills, and training. Third, tax policy has the potential to discourage productivity growth by attenuating research and development (R&D) and the development of venture capital for high-tech industries, activities whose spillover effects can potentially enhance the productivity of existing labor and capital. Fourth, tax policy can also influence the marginal productivity of capital by distorting investment from heavily taxed sectors into more lightly taxed sectors with lower overall productivity

(Harberger, 1962, 1966). And fifth, heavy taxation on labor supply can distort the efficient use of human capital by discouraging workers from employment in sectors with high social productivity but a heavy tax burden. In other words, highly taxed countries may experience lower values of and , which will tend to retard economic growth, holding constant investment rates in both human and physical capital (Engen and Skinner, 1992). We show this graphically in Figure 1, which focuses on a fixed level of the capital stock K, shown by the width of the horizontal axis. (A similar analysis holds for labor market distortions.) Suppose that the income tax on the corporate 2 sector, as well as subsidies to noncorporate owner-occupied housing, distort the allocation of the capital stock 2 between the corporate (c) and noncorporate (nc) sectors. (In other countries, the distortion may arise between sectors which escape taxation such as

FIGURE 1.

The Effect of Intersectoral Distortions on the Average Rate of Return

619

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

the underground economy or smallscale agriculture, versus the manufacturing sector which is easily taxed or heavily regulated.) The line denoted MP(c) is the value of the marginal product of capital in the corporate sector, while MP(nc) denotes the value of the marginal pro-duct in the noncorporate sector. Without any tax distortion, the profit-maximizing and most efficient point is C; the mar-ginal productivity of capital is equalized in both sectors and the economy-wide return on capital is R*, as shown by the dotted line. (The allocation of the total capital stock, K, is Q* units of capital in the noncorporate sector and K-Q* units in the corporate sector.) With a tax of AB on corporate capital only, there is a distortion in the allocation of capital; capital flows from the corporate to the noncorporate sector, so the new allocation is Q units of capital in the noncorporate sector and K-Q units of capital in the corporate sector. The net loss in output is given by ABC, the traditional Harberger welfare loss triangle. Under some plausible restrictions, the average rate of return for the entire capital stock, R, will correspond to the rate of return on new investment, given in equation 1 by .4 Hence, a distortionary tax on capital (or on labor) will be reflected in lower overall rates of return on new investment (from R* to R), leading to laggard growth rates. We have outlined five possible mechanisms by which taxes can affect economic growth. Therefore, it might appear that taxes should play a central role in determining long-term growth. However, the conventional Solow growth model implies that taxes should have no impact on long-term growth rates. In part, this result occurs by assumption, since productivity growth is assumed to be fixed and unaffected by tax policy. But this paradoxical result holds also
620

because of a distinction between changes in the level of GDP and changes in growth rates of GDP. For example, suppose that, in the year 2000, a tax and spend president is elected in the United States and tax rates are increased 3 by ten percentage points across the board. (Ignore the effects of the extra government spending on the economy.) The extra tax distortion reduces labor supply and investment, causing a sudden decline in short-term growth rates. But once the U.S. economy had adjusted to the harsh new tax regime, it would revert back to its original growth path, albeit at a lower absolute level than it would have been in the absence of the tax hikes. (In the Solow model, both investment and labor supply growth revert back to their original rates determined by long-term population growth.) In other words, the simple Solow model implies that tax policy, however distortionary, has no impact on long-term economic growth rates, even if it does reduce the level of economic output in the long-term. So then how can taxation affect output growth rates? We focus on two possible mechanisms. The first is that when the structure of taxes changes, short-term output growth rates would be expected to change as well along a possibly lengthy transition path to the new steady state. If one believes that the Dole or the Forbes tax reform would expand output by five percentage points and it takes ten years to make the transition to the new steady state, growth rates will be higher, on average, by about 0.5 percentage points during this period before settling back down to their long-run values.5 Ten years is a long-term horizon for presidential candidates but is still the short-term in the Solow model. And these short-term effects are clearly important, since they result in a permanent increase in GDP.

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

The second possibility arises within the context of the new class of endogenous growth models (e.g., Romer, 1986; Lucas, 1990). In these models, the stable growth rate of the Solow model, stapled down by technology and workforce productivity growth, is replaced by steady-state growth rates which can differ, persistently, because of tax and expenditure policies pursued by the government (e.g., King and Rebelo, 1990). The endogenous growth framework emphasizes factors such as spillover effects and learning by doing, by which firm-specific decisions to invest in capital or in R&D, or individual investments in human capital, can yield positive external effects (e.g.,on ) that benefit the rest of the economy. In these models, taxes can then have long-term, persistent effects on output growth. However, the question still remains: what is the magnitude of these tax effects on economic growth? A number of recent theoretical studies have used endogenous growth models to simulate the effects of a fundamental tax reform on economic growth.6 All of these studies conclude that reducing the distorting effects of the current tax structure would permanently increase economic growth. Unfortunately, the magnitude of the increase in economic growth is highly sensitive to certain assumptions embodied in the economic models used in these studies, with little empirical guidance or consensus about key parameter values. Consequently, these studies reached substantially different conclusions concerning the magnitude of the boost in growth rates. At one extreme, Lucas (1990) calculated that a revenue-neutral change that eliminated all capital income taxes while raising labor income taxes would increase growth rates negligibly. At the other extreme, Jones, Manuelli, and
621

Rossi (1993) calculated that eliminating all distorting taxes would raise average annual growth rates by a whopping four to eight percentage points.7 (An across-the-board reduction in distortionary tax rates in these models, rather than complete elimination of distortionary taxes, would be expected to have a smaller positive effect on economic growth.) Most recently, the simulation model in Mendoza, Razin, and Tesar (1994) suggests relatively modest differences in economic growth of roughly 0.25 percentage points annually as the consequence of a 10 percentage point change in tax rates. These simulation models of endogenous growth fail to provide a comfortable range of plausible effects of taxes on growth and thus tend to raise more questions than they answer. Moreover, they are likely to miss many relevant characteristics of the U.S. tax system. No macroeconomic model allows for the possibility of a firm undertaking financial restructuring to reduce taxable income, or of timing issues in deferred taxes, or the possibility of tax evasion.8 Often the simulation analysis is performed in terms of a single flat-rate tax in the context of a (single) representative agent model. Ultimately, one needs to consider the empirical record to make informed judgments about whether tax policy exerts a strong influence on economic growth. Below, we take three separate approaches to judge the empirical record. First, we take a quick look at the U.S. historical record to see if there is an easily discernible link between changes in U.S. tax policy and changes in economic growth across time. Second, we consider whether differences in growth rates across countries can be attributed, at least partially, to variation in tax policy. Third and finally, we survey

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

the microlevel studies of how taxes affect specific subsectors of the economy and build up from these microlevel studies to make inferences about aggregate tax effects. AN INFORMAL LOOK AT TAXES AND U.S. ECONOMIC GROWTH Anecdotal stories about the U.S. tax code can sometimes have a larger impact on the policy debate than a stack of statistical studies. The Kemp Commission (NCR, 1996), for example, highlighted the complaint of one frustrated businessman: As an entrepreneur, I experience first hand the horrors of our tax system. It has grown into a monstrous predator that kills incentives, swallows time, and chokes the hopes and dreams of many. We have abandoned several job-creating business concepts due to the tax complexities that would arise. While this testimony is suggestive that the tax system adversely affects incentives, it is not entirely clear whether the entrepreneur is concerned about the tax rate per se or the complexity of the tax system more generally. And we are not sure what fraction of entrepreneurs are of like mind, or how much investment is affected adversely by the tax code. For example, surveys from a few decades ago indicate that typical businesspeople did not view taxes as an impediment to business decisions; in one study conducted in Britain in the early 1960s, not a single executive out of the sample of 181 replied that they abandoned the introduction of a new plant or equipment during the past seven years because of tax changes (Corner and Williams, 1965).9 More recent survey studies suggest a larger impact of taxation on the discount rates used to evaluate private investment projects (Poterba and Summers, 1995); even
622

among these tax-savvy Fortune 1000 executives, 36 percent reported that a corporate tax cut from 34 to 25 percent would not make them more likely to engage in investment projects.10 A slightly more rigorous approach is to look at the historical evidence from time-series changes in taxation and output growth. The Kemp Commissions report (NCR, 1996) relied on time-series comparisons to argue that the patterns are self-evident: America has experienced three periods of very strong economic growth in this century: the 1920s, the 1960s, and the 1980s. Each of these growth spurts coincided with a period of reductions in marginal tax rates. In the eight years following the HardingCoolidge tax cuts, the American economy grew by more than five percent per year. Following the Kennedy tax cuts in the early 1960s, the economy grew by nearly five percent per year. . . In the seven years following the 1981 Reagan tax cuts, the economy grew by nearly four percent per year while real federal revenues rose by 26 percent. This approach does not try to perform the growth accounting exercise detailed in the theoretical section, but asks simply whether there are discernible differences in GDP growth following tax cuts. We consider the latter two tax reforms in Figure 2, which shows real GDP growth rates (both total and per capita) in the United States between 1959 and 1994 in the bottom panel, with the relevant tax series graphed in the upper two panels.11 To smooth out year-to-year volatility in GDP growth rates, we present three-year moving averages of GDP growth rates in the bottom panel of Figure 2, both for aggregate growth rates and for per capita growth rates. The two economic expansions noted above during the 1960s and the 1980s are apparent, as

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

FIGURE 2.

Average Tax Rates, Marginal Tax Rates, and GDP Growth in the United States, 195995

623

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

are the other expansions following recessions (shown by the shaded regions). The general slowdown in economic growth over the last three decades can be seen also. Moving to the top panel of Figure 2, we next consider the ratio of tax revenue to GDPa commonly used measure of the average tax burden. The top line shows U.S. federal government revenue (measured on a National Income and Product Accounts (NIPA) basis) as a percentage of GDP. The lower line is state and local government tax revenue measured on a NIPA basis as a percentage of GDP. Since 1959, the average federal tax rate has risen by about two percentage points, but has generally hovered around 20 percent of GDP; the average individual income tax rate has remained relatively constant, while growth in social insurance taxes have been mostly offset by the decline in corporate and excise taxes. State and local government average tax burdens have risen by about three percentage points over the last three decades. The KennedyJohnson tax cuts in 1964 resulted in a small decline in the average tax rate. Real GDP growth averaged a robust 4.8 percent over the subsequent 1964 to 1969 period. However, the extent to which this growth was caused by the tax cuts is unclear, as GDP growth had averaged over five percent in the two years prior to 1964. The Reagan tax cuts also lowered the average tax rate, and real GDP growth averaged a healthy 3.9 percent from 1983 to 1989, significantly above the preceding period from 1980 to 1982 that was dominated by recession.12 But it is a difficult task to sort out whether the strong growth during the 1980s was the consequence of supply-side effects

of lowering marginal tax rates, traditional Keynesian aggregate demand effects fueled by tax cuts and expanding defense expenditures, or a recovery that would have occurred without the tax change.13 Indeed, Feldstein and Elmendorf (1989) suggest a somewhat different cause for the 1980s expansion: expansionary monetary policy combined with a strong dollar and active business investment. Over the longer term, since 1959, both the average federal tax rate and the average state-local tax rate have risen by about two and three percentage points, respectively. At the same time, average growth rates in real GDP have declined, from 4.4 percent during the 1960s to only 2.4 percent in 198695. These coincident trends over the last three and a half decades are consistent with the hypothesis that higher taxes have stunted economic growth. Before arriving at conclusions about taxation and growth from this single observation (which does not account for other factors that were also changing over this time period), we note that the average tax rate series is unlikely to reflect the marginal tax distortion, which economic theory suggests is more important in affecting economic growth through households and firms choices of saving, investment, and employment. The middle panel of Figure 2 shows the marginal individual income tax rates relevant for households at the 75th, 50th, and 25th percentiles of the income distribution in each year (Hakkio, Rush, and Schmidt, 1996).14 From 1960 to the early 1980s, marginal tax rates at the 75th percentile grew while marginal tax rates at the 25th percentile declined slightly. There was some reduction in output growth coincident with the increase in the

624

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

upper-middle class marginal tax rates. However, GDP growth rates continued to fall over the past decade even as the marginal tax rates for both upper- and lower-income households declined.15 In other words, the time-series correlation between marginal tax rates and growth rates yields a decidedly mixed picture; some decades were correlated positively, and others negatively. Finally, we correct the first sentence of the quotation from the Kemp Commission above. The most rapid growth rates in this century were, in fact, during the period 194045, when output grew at 12.5 percentage points annually. During this same period, the federal tax system expanded dramatically, with median marginal tax rates rising from 3.6 percent in 1940 to 25 percent in 1945. Yet it would be ludicrous to claim on that basis that higher taxes have a positive effect on output growth, given the obvious confounding events during this period. Nevertheless, highlighting the period 194045 is useful for two purposes. The first is that it illustrates the risks of trying to discern incentive effects of taxation using short-term time-series data. This is a point reinforced by the experience of Swedens tax reform, when the economy fell into a recession just after a tax reform trimming marginal tax rates substantially (Agell, Englund, and Sodersten, 1996). And second, it suggests that one should look most carefully at GDP growth rates before and after the early 1940s when the federal income tax experienced its major expansion. Stokey and Rebelo (1995) looked for this break in longterm output growth rates and were unable to find any significant difference. On the other hand, given the major disruptions in economic activity occurring during the 20th century, it may be asking too much of the data to detect

what might be very small differences in growth rates, on the order of 0.5 percentage points, caused by the distortionary effects of taxation. More formal econometric methods may hold greater promise for uncovering the pure effects of taxation on economic growth, because that type of analysis attempts to control for other factors that affect output independently of tax policy. The problem is that time-series analysis is best suited for detecting short-term effects of changes in tax policy on output growth, which, as noted above, may reflect Keynesian expansionary effects of deficit spending or other unmeasured factors associated with tax cuts. In addition, figuring out which characteristics of a particular tax reformchanges in top marginal tax rates, depreciation allowances, tax progressivity, tax rates on capital gains caused changes in growth rates is particularly problematic in aggregate time-series analysis. For these reasons, we turn our attention next to crosscountry studies. TAX POLICY AND GROWTH: THE CROSSCOUNTRY EVIDENCE An alternative empirical approach is to draw on the experience of different countries to investigate how tax policy affects economic growth. Countries have very different philosophies about taxation and very different methods of collecting their revenue. During the past several decades, some countries have increased taxation quite dramatically, while, in other countries, tax rates have remained roughly the same. Some countries incorporated value-added taxation in the 1960s (e.g., France and Britain), while others shifted away from corporate taxation (the United States). The advantage of using such cross-

625

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

country comparisons is that we can use many countries with different tax structures and GDP growth rates to test for correlation (and, one hopes, causation) between tax policy and growth. In general, studies of taxation using cross-country data suggest that higher taxes have a negative impact on output growth, although these results are not always robust to the tax measure used. Using reduced-form cross-section regressions, Koester and Kormendi (1989) estimated that the marginal tax rateconditional on fixed average tax rateshas an independent, negative effect on output growth rates. Skinner (1988) used data from African countries to conclude that income, corporate, and import taxation led to greater reductions in output growth than average export and sales taxation. Dowrick (1992) also found a strong negative effect of personal income taxation, but no impact of corporate taxes, on output growth in a sample of Organisation for Economic Co-operation and Development (OECD) countries between 1960 and 1985. Easterly and Rebelo (1993) found some measures of the tax distortion (such as an imputed measure of marginal tax rates) to be correlated negatively with output growth, although other measures of the tax distortion were insignificant in the growth equations. Most empirical studies of taxation and growth are reduced form estimates in that they specify a linear model of output growth rates, with tax rates, labor resource growth, and investment rates on the right-hand side of the equation. However, taxes do not necessarily enter the growth accounting framework in equation 1 in a linear fashion. We explored this possibility in Engen and Skinner (1992), where the primary growth effect of tax distortions on production is hypothesized to
626

depress the economy-wide return on capital, , and on labor, (as in equation 1 and Figure 1). Using crosscountry data for 197085, Engen and Skinner found that an increase of 2.5 percentage points in the average tax burden (total taxes divided by GDP) is predicted to reduce long-term output growth rates by 0.18 percentage points, holding constant the supply of investment and labor. A recent McKinsey (1996) study points to the potential importance of the intersectoral allocation of capital. The study observed that Japan and Germany both had much higher rates of investment. But because U.S. investment appeared to be allocated to more profitable (i.e., higher productivity) sectors, the net increment to the effective capital stock, and hence to national income, was considerably greater in the United States, despite the lower investment rate. Similarly, King and Fullerton (1984), in their study of tax systems in the United Kingdom, Sweden, West Germany, and the United States, found a strong negative correlation between economic growth and the intersectoral variability in investment tax rates.16 Of course, nearly any tax will tend to distort economic behavior along some margin, so the objective of a welldesigned tax system is to avoid highly distortionary taxes and raise revenue from the less distortionary ones. There is some evidence that how a country collects taxes matters for economic growth. Figure 3, reproduced from Mendoza, Milesi-Ferretti, and Asea (1996), shows the correlation among the OECD countries between income taxes and economic growth (panels A and B) and consumption taxes and economic growth (panel C), over the period from 1965 to 1991. These scatter

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

FIGURE 3A.

Growth and the Capital Income Tax, OECD Countries

plots, largely confirmed in regression analysis, suggest that income taxation is more harmful to growth than broadbased consumption taxes. It is useful to consider the growth effects of a major tax reform using these cross-country regression estimates. Suppose that marginal tax rates are cut by a uniform five percentage points and average tax rates are cut by 2.5 percent of GDP, leading to a (static) revenue loss of $185 billion annually. This hypothetical tax reform was chosen because it is on the outer fringe of politically feasible tax reform, losing more than twice as much revenue as the tax proposal supported by presidential candidate Robert Dole. Were such a plan enacted, the tax-to-GDP ratio would revert to levels last seen in 1958. As noted above, the estimated coefficient from Engen and Skinner (1992) that ignores possible changes in the supply of capital and labor implies an increase in long-term
627

growth rates of 0.18 percentage points. Including estimates of the responsiveness of investment to the marginal tax rate from Mendoza, Milesi-Ferretti, and Asea (1996) suggests that this hypothetical tax reduction would increase investment by 1.35 percent, boosting the predicted growth rate effect of the tax cut to 0.32 percentage points annually.17

SANDTRAPS IN CROSS-COUNTRY ECONOMETRIC ANALYSIS To this point, we have been taking the results of the cross-country econometric studies at face value. Any empirical study must be treated with some caution; but, in many of the studies cited above, particularly the crosscountry studies, one must be particularly careful in the interpretation of the coefficients (Levine and Renelt, 1992; Slemrod, 1995). We consider just four of these potential problems below.

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

FIGURE 3B.

Growth and the Labor Income Tax, OECD Countries

FIGURE 3C.

Growth and the Consumption Tax, OECD Countries

628

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

First, studies of taxation and growth may find negative growth effects resulting from taxation, but it is more difficult to measure the potential benefits of the spending financed by the revenue collected. The combined impact of distortionary taxes and beneficial government expenditures may yield a net improvement in the workings of the private sector economy (e.g., Barro, 1990, 1991a,b). An example of the deleterious effects caused by the absence of government spending comes from the World Development Report (World Bank, 1988, p.144): According to the Nigerian Industrial Development Bank (NIDB), frequent power outages and fluctuations in voltage affect almost every industrial enterprise in the country. To avoid production losses as well as damage to machinery and equipment, firms invest in generators.. . . One large textile manufacturing enterprise estimates the depreciated capital value of its electricity supply investment as $400 per worker.. . . Typically, as much as 20 percent of the initial capital investment for new plants financed by the NIDB is spent on electric generators and boreholes. That is, when the government of Nigeria did not provide the necessary electricity supply, private firms were forced to generate electricity on their own, and presumably at much higher cost. Clearly, a tax in Nigeria earmarked for (new) government expenditures on improving the electrical system would be likely to enhance economic growth even if the taxes distorted economic activity. The problem is that taxes are not necessarily earmarked to those expenditures most conducive to economic growth, either because of political inefficiencies or because of redistributional policies that may yield benefits for society but will not be reflected in robust GDP growth rates (Atkinson, 1995).18 Thus, one
629

must be careful in interpreting the coefficients on tax and output growth studies to remember that these estimates reflect just one partthe costs of a combined tax and expenditure system. Second, one should be very wary of the data, particularly from developing countries with large agricultural or informal sectors where the measurement of income is difficult indeed.19 Even in developed countries, it is well known that GDP measures suffer from biases and mismeasurement of productivity in service sectors, for example.20 Measuring the effective tax rate is even more difficult, given the wide variety of tax distortions, methods for measuring them, and variation across countries in administrative practices. Third, there are real difficulties with reverse causation; one does not know whether regression coefficients reflect the impact of investment on GDP growth rates, for example, or the reverse influence of GDP growth rates on investment, or both effects combined (Blomstrom, Lipsey, and Zejan, 1996). Sometimes these biases creep in because of the way the regression variables are constructed. Suppose one wanted to estimate an explicitly shortterm relationship between the change in the tax burden, typically measured as the ratio of tax revenue to GDP, and the percentage growth rate in GDP. Any positive measurement error (or shortterm shock) in GDP will shift GDP growth rates up but also tend to shift the tax-to-GDP ratio down, thereby introducing a spurious negative bias in the estimated coefficient.21 One can try to avoid such bias by introducing as explanatory variables the percentage growth rate in the level of taxation, or of government expenditures, rather

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

than the change in the ratio, as above. In this case, the bias would go in the opposite direction, because countries that grow rapidly also tend to experience rapid growth in tax collection and in spending.22 One approach for both of these problems is to use instrumental variables for changes in government spending and taxation (Engen and Skinner, 1992), although the problem still remains to find appropriate exogenous instruments. Another reverse causality problem comes in deciding what factors to include on the right-hand side of a growth regression. Should one control for other factors such as inflation, political unrest, and the share of agriculture in total output? On the one hand, these are factors that could be spuriously correlated with tax policy, and one would clearly want to control for them. But, on the other hand, a shrinking share of agriculture in output, or political unrest, or inflation could be symptomatic of the underlying growth rate of the economy. During severe recessions, countries often resort to high inflation rates as a means of financing expenditures after their tax collection efforts have collapsed. This reverse causation makes it harder to argue that inflation causes poor economic growth, as well as making it difficult to interpret the coefficients on all other variables. In sum, reverse causality is really the Achilles heel of the typical cross-country regression. Nearly every variable on the right-hand side of the regression is suspect. Fourth, as noted by Slemrod (1995), countries may differ both in their tastes for government-sector spending (the demand side) and in their ability to raise tax revenue (the supply side). Suppose that more developed countries experience a lower cost of raising tax revenue,
630

perhaps because industrial production is much easier to tax than agricultural production. Then countries that grow quickly may also experience a more pronounced drop in their cost of raising tax revenue, which could in turn lead to more rapid growth in tax revenue. The researcher might well find a spurious positive correlation between tax rates and output growth. By the same token, countries that grow fast may exercise a greater taste for government spending (sometimes known as Wagners law), leading to a shift to the right in the demand for government spending. As Slemrod points out, such a model would imply that, in a cross section of countries, there could be little correlation between output growth, government spending, and taxation.23 Slemrods point is therefore a cautionary one, that the regression coefficients one actually estimates may have little to do with the Solow-style production function written in equation 1 (see also Islam, 1995). But this point also suggests that, even if taxes affect growth rates adversely, cross-country regression models would be biased against detecting such effects. SECTORAL STUDIES OF TAXATION AND GROWTH Our third approach is to consider separately the effect of taxes on the disaggregated micro components in equation 1, such as labor supply, human capital, investment, and technological growth. We then combine these effects to arrive at an aggegrate bottom-up measure of how our hypothetical tax reformcutting marginal tax rates by five percentage points, and average rates by 2.5 percentwill affect output growth.24 The advantage of this approach is a more accurate measure of how economic agents respond to tax incentives, often with data generated by natural experiments such as tax reform

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

or other (exogenous) legislative change. There are two disadvantages to this strategy, however. First, we are unable to account for the spillover effects of both human and physical capital accumulation, as in the hypothesized correlation between the level of investment and technological innovation (Boskin, 1988). And, second, even with this disaggregated approach, there is virtually no empirical evidence on some key parameter values. Change in the Labor Force Consider first the effects of taxation on labor supply. The top panel of Figure 4 contains a graph that shows marginal labor income tax rates for the United States from 1965 to 1988 from Mendoza, Razin, and Tesar (1994) plotted against the average weekly hours for workers in private nonagricultural industries and also the civilian labor force participation rate. As labor income tax rates have increased, average weekly hours have declined. On the other hand, labor force participation has increased. (Although not shown, participation has generally increased for women while falling for men.) Thus, the effect of increased marginal labor taxes appears to be ambiguous based on this simple time-series examination. A voluminous empirical literature has examined how taxes affect the labor supply of individuals within various demographic groups (e.g., Killingsworth, 1983; Hausman, 1985; MaCurdy, Green, and Paarsch, 1990; Triest, 1990, 1996; Bosworth and Burtless, 1992; Mariger, 1995; Eissa, 1996a,b). Generally, the results suggest quite modest labor supply effects of tax policy in the United States.25 Most estimates suggest that both work hours and labor force participation for men are only mildly responsive to historically
631

experienced tax changes, and Heckman (1993) concludes that most of the evidence points to a relatively larger participation effect than hours effect. Estimated uncompensated tax elasticities are usually small, often in the range of zero to 0.1.26 Recently, Eissa (1995) found that married women in highincome households are more responsive to tax changeswith tax elasticities in the range of 0.6 to 1with approximately equal importance on hours and participation changes. However, working married women make up a relatively small part of the labor force and often have relatively tenuous ties to the labor force (Eissa, 1996a). Like men, unmarried women generally have similarly small labor supply responses to taxes (Eissa, 1996a). For the purposes of our equation 1, we would like to know how tax policy affects the rate of change in quality. adjusted labor supply m. Consider first short-term effects. If the labor supply elasticity is assumed to be 0.15 and marginal tax rates decline by five percentage points, then one might expect an increase of 0.75 percent in total hours worked. Assuming labor income comprises 75 percent of total output and the labor market transition is spread over a ten-year transition period, the net change in GDP growth rates over the short-term (ten-year) period would be 0.06 percent annually. In the long-term, however, only taxinduced changes in the accumulation of education or human capital more generally would affect the growth . rate m. A number of empirical studies (e.g. Romer, 1990; Mankiw, Romer, and Weil, 1992; Judson, 1996) suggest that measures of human capital have statistically and economically important effects on economic growth, although

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

FIGURE 4.

Labor, Investment, and Factor Tax Rates, 196588

632

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

some (e.g., Barro and Lee, 1992) estimate that the effect is quite small. However, the effect of taxes on human capital formation is quite uncertain. Theoretically, the effect is ambiguous and, not surprisingly, simulation analysis can lead to a variety of conclusions. Trostel (1993) simulates substantial long-term elasticities of human capital with respect to taxation; he suggests a long-term increase in human capital of 0.97 percent per one percentage point decrease in the marginal tax rate. Hence, our hypothetical five percentage point reduction in the marginal tax rate would be predicted to increase the stock of human capital by 4.8 percent. In equilibrium, maintaining that higher level of human capital requires an extra 4.8 percent additional net investment in . human capital. Suppose that mi were about three percent annually. The new level of equilibrium growth in human capital would rise to 3 1.048, or 3.14 percent annually.27 Assuming the factor share coefficient is 0.75, the net effect on growth would be 0.10 percentage points.

Milesi-Ferretti, and Asea (1996). There is a moderate negative correlation between tax rates and investment rates; more detailed regression analysis suggests that a 10 percentage point change in tax rates on profits could affect investment rates by at most two percentage points. It should be noted, however, that one shortcoming of these capital tax measures is that they use weighted statutory rather than effective rates, and thus they cannot account for the dramatic increase in effective marginal tax rates on capital during periods of inflation (e.g., King and Fullerton, 1984; Fullerton and Karayannis, 1993). A number of recent studies (e.g., Auerbach and Hassett, 1991; Cummins, Hassett, and Hubbard, 1994, 1996; Chirinko, Fazzari, and Meyer, 1996) have found significant effects of tax policy on investment, suggesting a plausible range for the investment elasticity for changes in the user cost of capital in the range of 0.25 to 1. This finding is potentially important because, although Levine and Renelt (1992) find that almost all results are fragile in cross-country growth regressions, they do find a positive, robust correlation between growth and investment. How might a change in the nature of investment decisions affect output growth? Suppose we adopt an investment elasticity of 0.5; then, a five percentage point drop in marginal tax rates should boost investment rates by 2.5 percent, or by about 0.4 percent of GDP. Assuming the net marginal product of capital is ten percent, output growth rates might be expected to grow by another 0.04 percentage points. We assume this boost in growth rate will be permanent, although in the Solow-style model, the growth effects will diminish over time.
633

Change in the Net Investment Rate The bottom panel of Figure 4 shows marginal capital income tax rates for the United States from 1965 to 1988 from Mendoza, Razin, and Tesar (1994), plotted against private nonresidential fixed investment as a percentage of GDP. As has been noted before (e.g., Chirinko, 1993; Hassett and Hubbard, 1996), a simple examination of the timeseries evidence suggests little relationship (and possibly a positive correlation) between investment and capital income tax rates. However, as before, this type of analysis is surely too simplistic. Alternatively, Figure 5 shows a graph from data on the OECD countries comparing capital income taxes with investment rates, taken from Mendoza,

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

FIGURE 5.

Capital Income Taxation and Investment Rates, OECD Countries

One factor that could stifle tax-induced investment expansions is a lack of new saving to finance the increased investment. In an economy without foreign capital flows, the increased demand for investment would be financed by the additional supply of saving attracted by higher net interest rates. But simulation models (Engen, 1996) and empirical studies (Skinner and Feenberg, 1990) find little support for a strong responsiveness of personal saving to the interest rate (although, see Elmendorf, 1995, and references cited therein). The relevant source of financing for the extra investment may therefore be retained earnings of firms and foreign investors.28 In any case, the investment elasticities gained from microlevel studies of firm investment behavior already reflect the additional cost or difficulty incurred by firms in providing additional financing for their investments, suggesting that the pure demand elasticities are even larger.
634

The Impact of Taxation on the Productivity Residual Taxes can affect the output growth in another way, by discouraging innovations and economic organizations that result in increased levels of output, holding constant the supply of capital and labor. In other words, distortionary tax policy may permanently reduce the level of technological growth . Of course, by its nature, trying to determine whether the residual effect is caused by tax policy or by some other factor (of which there are always many candidates) is always problematic. Here, we consider two examples: the effects of tax policy on research and development and its impact on entrepreneurship. Hall (1993) studied the impact of the tax credit on R&D spending using two sources of variation: changes in the tax code over the 1980s and differences in the taxable status of individual firms

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

that affected their ability to take advantage of the credits. She found quite large effects: for every $1 billion lost in tax revenue, there was a $2 billion increase in R&D spending. Since R&D is about 2.5 percent of GDP (Nonneman and Vanhoudt, 1996), Halls estimates imply that a five percentage point tax advantage to R&D would increase R&D spending by 0.25 percent of GDP. Using a rate of return to R&D spending of 30 percent (e.g., Griliches, 1988), the net effect would be a 0.075 increase in GDP growth rates. A second possibility is that the hypothetical tax cut, for example, on capital gains, would stimulate entrepreneurship and innovation, which in turn would augment productivity growth. Poterba (1989) investigated the tax incentives faced by venture capitalists, and concluded that venture capital was only a small fraction of total capital income, so that tax cuts were a blunt sword to encourage hightech industries. Furthermore, tax-exempt institutions provided a large fraction of start-up funds, and these institutions are not subject to income taxation. A somewhat different picture emerges from a recent study quantifying labor hiring decisions by self-employed workers. Carroll et al. (1996) found that a six percentage point decline in the marginal tax rate of a (Schedule C) entrepreneur in the top tax bracket increased by 11 percent the likelihood of hiring at least one employee. However, the magnitude of these effects and their impact on aggregate employment are just not well enough understood to hazard a numerical estimate of their growth effects. Summing Up To complete our bottom-up analysis, we simply add the growth effects based on
635

changes in human capital, investment, and technological growth. The long-run effects of our hypothetical major tax reform are estimated to be 0.22 percentage points, while the short-term effects, which include the transitional effects of increased labor supply, increase to 0.28 percentage points. Aside from the uncertainty inherent in nearly every empirical parameter used in these calculations, there are some further caveats. First, the calculation ignores the reduction in the sectoral distortion of capital and labor, which, in the section on cross-country regressions, was found to be important. And second, these estimates reflect a uniform reduction of five percentage points in marginal tax rates for all income-generating activities. It may be the case that tax cuts in capital gains, or tax credits for R&D, coupled with increases in consumption taxes, or a shift to a flat tax, could yield much stronger growth effects with less pronounced revenue effects. Nevertheless, these results suggest growth effects from major tax reform on the order of one-quarter of one percent per year. LESSONS FOR POLICY While the last word on taxation and economic growth certainly has not been heard, there are some lessons that we think can be taken from the evidence thus far. First, we think that tax policy does affect economic growth. There is enough evidence linking taxation and output growth to make the reasonable inference that beneficial changes in tax policy can have modest effects on output growth. The implied effects from the bottom-up microlevel studies and the top-down cross-country regressions are quite close in magnitude: a

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

major tax reform reducing all marginal rates by five percentage points and average tax rates by 2.5 percentage points is predicted to increase longterm growth rates by between 0.2 and 0.3 percentage points. Whether these effects on output growth are permanent (lasting forever) or transitory (lasting perhaps 10 to 15 years) is difficult to determine, both because our data sources do not extend for a lengthy period and because tax regimes themselves generally have such short half-lives. Second, even these modest growth effects can have an important long-term impact on living standards. For example, suppose that an inefficient structure of taxation has, since 1960, retarded growth by 0.2 percent per annum. Accumulated over the past 36 years, the lower growth rate translates to a 7.5 percent lower level of GDP in 1996, or a net reduction in output of more than $500 billion annually. So the potential effects of tax policy, although difficult to detect in the time-series data, can be potentially very large in the long term. Third, it appears highly unlikely that past tax reforms have been self-financing in the aggregate. There is evidence that tax changes focused on high-income taxpayers may be self-financing, perhaps because of changes in financial arrangements as well as shifts in economic activity (e.g., Feldstein, 1995; Feldstein and Feenberg, 1996). Of course, the historical record does not relate specifically to a flat tax or a consumption-based tax, which could have quite different effects, but we think it unlikely that any tax system could engender the long-term increases in growth rates necessary to completely pay for the tax cuts. We want to be careful here about the context of our conclusions about
636

taxation and growth in the policy debate over dynamic scoring. Typically, dynamic scoring of tax revenue in response to changes in the tax code involves two adjustments: one is the microeconomic change in the tax base, holding constant macroeconomic variables, and the other is the change in macroeconomic climate caused by the tax reform (Auerbach, 1996b). Here, we say nothing about the first, microeconomic effects, which could well be quite large (as in the short-term response of capital gains realization to changes in the capital gains tax cut). We simply claim that the second, macroeconomic, effect is likely to be quite modest. Fourth, a major shortcoming with nearly all cross-country and time-series studies is the difficulty of measuring the marginal tax burden appropriately. The average tax rate does not reflect the marginal tax burdens hypothesized to affect economic decisions. Even statutory marginal tax rates may not adequately reflect the quite complex intertemporal incentive effects of a complex tax system. In many countries, tax policy is administered at the local level, where the tax collector may not even have a current copy of the relevant statutes. Fifth, the composition of the tax system is probably as important for economic growth as is the absolute level of taxation. Countries that are able to mobilize tax resources through broadbased tax structures with efficient administration and enforcement will be likely to enjoy faster growth rates than countries with lower overall tax collections assessed inefficiently. In short, the design of the tax system is likely to exert a modest, but cumulatively important, influence on long-term growth rates.

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

ENDNOTES
We are grateful for the very helpful comments from Darrel Cohen, Don Fullerton, William Gale, Kevin Hassett, Harvey Rosen, and Joel Slemrod. The opinions expressed in this paper are those of the authors and are not necessarily shared by the Board of Governors of the Federal Reserve System or other members of its staff.
1

11

2 3

10

The Kemp Commission was formally known as the National Commission on Economic Growth and Tax Reform (1996). For examples, see Gravelle (1995) and Gale (1996). The two coefficients are not measured in the same . units because ki is expressed as a ratio of GDP and . mi as a percentage change. See Auerbach, Hassett, and Oliner (1994) for a discussion of how corresponds to the (net or gross) return on capital. David (1977) suggests that much of the 19th century in the United States was characterized by a transition from a low to a high capital-intensity economy. On the other hand, King and Rebelo (1993) find that traditional Solow growth models generate implausible transition paths in shifting from one equilibrium to another. There is an extensive simulation literature showing transitional gains in economic efficiency using the framework of dynamic computable general equilibrium models; see Ballard et al. (1985), Auerbach and Kotlikoff (1987), Fullerton and Rogers (1993), Auerbach (1996a), and Engen and Gale (1996). Like the endogenous growth literature, the results from such studies often depend on the structure of the simulation model. In a life cycle model with perfect certainty and perfect foresight, Auerbach and Kotlikoff (1987) and Auerbach (1996a) find quite dramatic shifts in some aggregate variables (such as saving rate) during the transition to a new steady state. In a model with uncertainty about future earnings, Engen and Gale (1996) find more moderate shifts in output and saving during the transition to a new tax regime. Stokey and Rebelo (1995) provide an excellent survey of this literature and explain why the theoretical simulation models differ so dramatically in their implications for growth. For a discussion of these issues, see Alm (1996), Slemrod (1990, 1994, 1995), and Auerbach and Slemrod (1997). Moreover, only eight percent said they had even postponed investment. Also see Holland (1969) for survey evidence on the labor supply of highly paid executives. Specifically, the survey question asked whether the tax cut would reduce or increase the hurdle rate or the minimum rate of return required before approving internal corporate investments. 637

Including earlier periods is complicated by the fact that revised GDP figures are currently only available on a consistent basis from 1959. Also, Lindsey (1990) notes that the CoolidgeMellon cuts in the 1920s affected only the top quarter of households as most U.S. citizens paid no income tax during that time. 12 During 197179, the economy expanded at an annual average rate of 3.5 percent including the recession years of 197475. Growth averaged 3.2 percent over the 198289 period. 13 Sorting out the difference between supply-side and demand-side expansions is important, since demand-side expansions tend to deflate later into recessions, while supply-side shifts correspond to permanent improvements in the productive capacity of the economy. 14 We are grateful to the authors for making this data on tax rates at different income percentiles available to us. Note that these tax rates only reflect the federal individual income tax and do not incorporate federal corporate income, earned income tax credit, payroll, or state income taxes. 15 An alternative measure of the tax distortion is the top statutory federal income tax rate. The top rate reached its zenith during the 1950s and early 1960s, when it was 91 percent. Since then it has bounced steadily downward to 28 percent, briefly, in 1988, with a jump back to 39.6 percent by 1993. (See Pechman, 1987, for a historical summary of most of this time period.) The economic expansion of the 1980s coincided with a marked decline in top marginal tax rates, leading some to conjecture a causal relationship between the cuts in top marginal rates and the economic expansion. However, taking the long view (circa 196088), a general decline in the top marginal rate occurred as average GDP growth rates tended to fall. 16 In the King and Fullerton study, based on 1980 data, West Germany exhibited the least degree of intersectoral distortion, trailed closely by the United States. In the McKinsey study, the factors identified as importantthe motivation of managers to show profits, for exampleare extremely difficult to quantify across countries on a consistent basis. Furthermore, as Kevin Hassett pointed out to us, the productivity of the capital stock may not necessarily be an indicator of better organization. In the absence of perfect world capital markets, a country may exhibit a higher productivity of capital because capital is scarce (that is, the capital-labor ratio is low). In this case, an increase in the capital stock might lower capital productivity but make the economy better off. 17 The investment effect is calculated using the first equation from Table 4 in Mendoza, Milesi-Ferretti, and Asea (1996), assuming that marginal labor and capital taxes are both cut by five percentage

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

points, while the output effect assumes a marginal product of capital equal to 0.10 (Auerbach, Hassett, and Oliner, 1994.) Unfortunately, we have no estimates from cross-country equations on labor supply effects. 18 Empirical evidence from a cross section of states suggests either that government spending yields no positive growth effects (Holtz-Eakin, 1994) or that only educational spending yields positive effects (Evans and Karras, 1994). Aschauer (1989) argues that the productivity effects are quite large. 19 The commonly used Summers and Heston (1991) data include a grade, ranging from A to D, that summarizes the authors estimate of the reliability of the data. Engen and Skinner (1992) weighted their estimates with a numerical scale of this reliability; results were similar, although standard errors were smaller. 20 For a nontechnical discussion, see The Real Truth About the Economy: Are Government Statistics Just So Much Pulp Fiction? (Business Week, November 7, 1994). 21 For example, Grier and Tullock (1989) find a negative correlation between output growth and the growth of government expenditures, although they do not interpret the correlation as reflecting reverse causation. 22 Rams (1986) estimated positive correlation between the growth in government spending and output growth appears to be an example of this problem. 23 The analogy is to market prices for competitive goods; regressing price on quantity (or conversely) tells the researcher nothing about the nature of the supply curve or of the demand curve without further identifying variables. 24 This is the approach followed by Agell, Englund, and Sdersten (1996) in considering the Swedish tax reform of the early 1990s. The bottom-up and top-down terminology is attributed to Slemrod (1995). 25 Lindsey (1987), Navratil (1994), Auten and Carroll (1995), Feldstein (1995), and Slemrod (1996) find evidence of behavioral responses to tax reforms by documenting increases in reported taxable incomes following reductions in tax rates during the 1980s. However, it is difficult in these analyses of taxpayers income to separate the effects of real responsessuch as changes in labor supplyfrom the effects of compensation, timing, and reporting responses. 26 We focus here on uncompensated elasticities, because we are considering a tax cut. However, if government expenditures are highly substitutable with market consumption goods or if Ricardian equivalence holds, one might prefer to use compensated elasticities, which are generally higher. 27 Strictly speaking, in the growth accounting framework in equation 1, the percentage growth . rate mi will be unaffected by the higher level of human capital because human capital growth is 638

28

defined in percentage terms. We instead consider an alternative renormalization in which the denominator is the pre-tax-cut level of human capital. Government tax policy could also be used to encourage saving through targeted saving programs such as IRAs or 401(k)s. While there is some debate about their effectiveness in increasing saving (see the Fall 1996 issue of the Journal of Economic Perspectives), the macroeconomic effects of these programs are probably not large given their modest size relative to GDP.

REFERENCES
Agell, Jonas, Peter Englund, and Jan Sdersten. Tax Reform of the Centurythe Swedish Experiment. National Tax Journal 49 No. 4 (December, 1996): 64162. Alm, James. What is an Optimal Tax System? National Tax Journal 49 No. 1 (March, 1996): 11733. Aschauer, David Alan. Is Public Expenditure Productive? Journal of Monetary Economics 23 No. 2 (March, 1989): 177200. Atkinson, Anthony B. The Welfare State and Economic Performance. National Tax Journal 48 No. 2 (June, 1995): 17198. Auerbach, Alan. Tax Reform, Capital Allocation, Efficiency, and Growth. In Economic Effects of Fundamental Tax Reform, edited by Henry Aaron and William Gale. Washington, D.C.: The Brookings Institution, 1996a. Auerbach, Alan. Dynamic Revenue Estimation. Journal of Economic Perspectives 10 No. 1 (Winter, 1996b): 14158. Auerbach, Alan, and Kevin Hassett. Recent U.S. Investment Behavior and the Tax Reform Act of 1986: A Disaggregate View. CarnegieRochester Conference Series on Public Policy 35 (Autumn, 1991): 185215. Auerbach, Alan, Kevin Hassett, and Stephen Oliner. Reassessing the Social Returns to Equipment Investment. Quarterly Journal of Economics 109 No. 3 (August, 1994): 789802. Auerbach, Alan, and Laurence Kotlikoff. Dynamic Fiscal Policy. Cambridge: Cambridge University Press, 1987. Auerbach, Alan, and Joel Slemrod. The Economic Effects of the Tax Reform Act of 1986. Journal of Economic Literature 35 No. 1 (forthcoming, March, 1997). Auten, Gerald, and Robert Carroll. Taxpayer Behavior and the 1986 Tax Reform Act. U.S. Treasury Department, Office of Tax Analysis. Mimeo, 1995.

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

Ballard, Charles L., Don Fullerton, John Shoven, and John Whalley. A General Equilibrium Model for Tax Policy. Chicago: The University of Chicago Press, 1985. Barro, Robert J. Government Spending in a Simple Model of Endogenous Growth. Journal of Political Economy 98 No. 5 (October, 1990, part 2): S10325. Barro, Robert J. Economic Growth in a CrossSection of Countries. Quarterly Journal of Economics 104 (May, 1991a): 40744. Barro, Robert J. A Cross-Country Study of Growth, Saving and Government. In National Saving and Economic Performance, edited by B. Douglas Bernheim and John Shoven. Chicago: University of Chicago Press and National Bureau of Economic Research, 1991b. Barro, Robert J., and Jong-Wha Lee. International Comparisons of Educational Attainment, 19601985. Harvard University. Mimeo, 1992. Blomstrom, Magnus, Robert E. Lipsey, and Mario Zejan. Is Fixed Investment the Key to Economic Growth? Quarterly Journal of Economics 111 No. 1 (February, 1996): 26976. Boskin, Michael. Tax Policy and Economic Growth: Lessons from the 1980s. Journal of Economic Perspectives 2 No. 4 (Fall, 1988): 71 97. Bosworth, Barry, and Gary Burtless. Effects of Tax Reform on Labor Supply, Investment, and Saving. Journal of Economic Perspectives 6 No. 1 (Winter, 1992): 326. Carroll, Robert, Douglas Holtz-Eakin, Mark Rider, and Harvey S. Rosen. Income Taxes and Entrepreneurs Use of Labor. Working Paper No. 32, Center for Economic Policy Studies, Princeton: Princeton University, July, 1996. Chirinko, Robert S. Business Fixed Investment Spending: Modeling Strategies, Empirical Results, and Policy Implications. Journal of Economic Literature 31 No. 4 (December, 1993): 18751911. Chirinko, Robert S., Steven M. Fazzari, and Andrew P. Meyer. Business Investment and the User Cost of Capital: New Evidence from U.S. Panel Data. Emory University, Atlanta, GA. Mimeo, 1996. Corner, D. C., and Alan Williams. The Sensitivity of Business to Initial and Investment Allowances. Economica 32 No. 125 (February, 1965): 3247. Cummins, Jason, Kevin Hassett, and R. Glenn Hubbard. A Reconsideration of Investment Behavior Using Tax Reforms as 639

Natural Experiments. Brookings Papers on Economic Activity 2 (1994): 174. Cummins, Jason, Kevin Hassett, and R. Glenn Hubbard. Have Tax Reforms Affected Investment? In Tax Policy and the Economy, Vol. 9, edited by James Poterba. Cambridge, MA: MIT Press, 1995. Cummins, Jason, Kevin Hassett, and R. Glenn Hubbard. Tax Reforms and Investment: A Cross-Country Comparison. Journal of Public Economics (forthcoming, 1996). David, Paul. Invention and Accumulation in Americas Economic Growth: A Nineteenth Century Parable. In Industrial Organization, National Priorities, and Economic Development, edited by Karl Brunner and Allan Meltzer. Amsterdam: NorthHolland, 1977. Dowrick, Steve. Estimating the Impact of Government Consumption on Growth: Growth Accounting and Optimizing Models. Australian National University. Mimeo, October, 1992. Easterly, William, and Sergio Rebelo. Fiscal Policy and Economic Growth: An Empirical Investigation. Journal of Monetary Economics 32 No. 3 (December, 1993): 41758. Eissa, Nada. Taxation and Labor Supply of Married Women: The Tax Reform Act of 1986 as a Natural Experiment. NBER Working Paper No. 5023. Cambridge, MA: National Bureau of Economic Research, 1995. Eissa, Nada. Tax Reforms and Labor Supply. In Tax Policy and the Economy, Vol. 10, edited by James Poterba. Cambridge, MA: MIT Press, 1996a. Eissa, Nada. Labor Supply and the Economic Recovery Act of 1981. In Empirical Foundations of Household Taxation, edited by Martin Feldstein and James Poterba. Chicago: University of Chicago Press, 1996b. Elmendorf, Douglas. The Response of Household Saving to Changes in Interest Rates. Federal Reserve Board, Washington, D.C. Mimeo, 1995. Engen, Eric. Precautionary Saving and the Structure of Taxation. Federal Reserve Board, Washington, D.C. Mimeo, 1996. Engen, Eric, and William G. Gale. The Effects of Fundamental Tax Reform on Saving. In Economic Effects of Fundamental Tax Reform, edited by Henry Aaron and William Gale. Washington, D.C.: The Brookings Institution, 1996. Engen, Eric, and Jonathan Skinner. Fiscal Policy and Economic Growth. NBER Working Paper No. 4223. Cambridge, MA: National Bureau of Economic Research, 1992. Evans, Paul, and Georgios Karras. Are Government Activities Productive? Evidence

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

from a Panel of U.S. States. The Review of Economics and Statistics 76 No. 1 (February, 1994): 111. Feldstein, Martin. The Effect of Marginal Tax Rates on Taxable Income: A Panel Study of the 1986 Tax Reform Act. Journal of Political Economy 103 No. 3 (June, 1995): 55172. Feldstein, Martin, and Douglas W. Elmendorf. Budget Deficits, Tax Incentives, and Inflation: A Surprising Lesson from the 198384 Recovery. In Tax Policy and the Economy 3, edited by Lawrence Summers. Cambridge, MA: MIT Press, 1989. Feldstein, Martin, and Daniel Feenberg. The Effect of Increased Tax Rates on Taxable Income and Economic Efficiency: A Preliminary Analysis of the 1993 Tax Rate Increases. In Tax Policy and the Economy 10, edited by James Poterba. Cambridge, MA: MIT Press, 1996. Fullerton, Don, and Marios Karayannis. United States. In Tax Reform and the Cost of Capital: An International Comparison, edited by Dale Jorgenson and Ralph Landau. Washington, D.C.: The Brookings Institution, 1993. Fullerton, Don, and Diane Lim Rogers. Who Bears the Lifetime Tax Burden? Washington, D.C.: The Brookings Institution, 1993. Gale, William. The Kemp Commission and the Future of Tax Reform. Tax Notes 70 No. 6 (February 5, 1996): 71729. Gravelle, Jane. The Flat Tax and Other Proposals: Who Will Bear the Tax Burden? Tax Notes 69 No. 12 (December 18, 1995): 151727. Grier, Kevin B., and Gordon Tullock. An Empirical Analysis of Cross-National Economic Growth. Journal of Monetary Economics 24 No. 2 (September, 1989): 25976. Griliches, Zvi. Productivity Puzzles and R&D: Another Nonexplanation. Journal of Economic Perspectives 2 No. 4 (Fall, 1988): 922. Hakkio, Craig, Mark Rush, and Timothy Schmidt. The Marginal Income Tax Rate Schedule from 1930 to 1990. Journal of Monetary Economics (forthcoming, 1996). Hall, Bronwyn. R&D Tax Policy During the 1980s: Success or Failure? In Tax Policy and the Economy 7, edited by James Poterba. Cambridge, MA: MIT Press, 1993. Hall, Robert, and Alvin Rabushka. The Flat Tax. 2d ed. Stanford: The Hoover Institution Press, 1995. Harberger, Arnold C. The Incidence of the Corporation Income Tax. Journal of Political Economy 70 No. 3 (June, 1962): 21540. 640

Harberger, Arnold C. Efficiency Effects of Taxes on Income From Capital. In Effects of Corporation Income Tax, edited by Marian Krzyzaniak. Detroit: Wayne State University Press, 1966. Hassett, Kevin A., and R. Glenn Hubbard. Tax Policy and Investment. NBER Working Paper No. 5683. Cambridge, MA: National Bureau of Economic Research, 1996. Hausman, Jerry. Taxes and Labor Supply. In Handbook of Public Economics, edited by Alan Auerbach and Martin Feldstein. Amsterdam: Elsevier, 1985. Heckman, James. What Has Been Learned About Labor Supply in the Past Twenty Years? American Economic Review 83 (May, 1993): 11621. Helms, L. Jay. The Effect of State and Local Taxes on Economic Growth: A Time Series-Cross Section Approach. Review of Economics and Statistics 67 No. 4 (November, 1985): 57482. Holland, Daniel. The Effect of Taxation on Effort: Some Results for Business Executives. In Proceedings of the Sixty-Second Annual Conference on Taxation, 428517. Columbus: National Tax AssociationTax Institute of America, 1969. Holtz-Eakin, Douglas. Public Sector Capital and the Productivity Puzzle. Review of Economics and Statistics 76 No. 1 (February, 1994): 1221. Islam, Nazrul. Growth Empirics: A Panel Data Approach. Quarterly Journal of Economics 110 No. 4 (November, 1995): 112770. Jones, Larry E., Rodolfo E. Manuelli, and Peter E. Rossi. Optimal Taxation in Models of Endogenous Growth. Journal of Political Economy 101 No. 3 (June, 1993): 485517. Judson, Ruth. Do Human Capital Coefficients Make Sense? A Puzzle and Some Answers. Federal Reserve Board, FEDS Working Paper No. 96-13. Washington, D.C.: Finance and Economic Discussion Series, 1996. Killingsworth, Mark. Labor Supply. Cambridge: Cambridge University Press, 1983. King, Mervyn, and Don Fullerton, eds. The Taxation of Income from Capital: A Comparative Study of the United States, United Kingdom, Sweden, and West Germany. Chicago: University of Chicago Press, 1984. King, Robert G., and Sergio Rebelo. Public Policy and Economic Growth: Developing Neoclassical Implications. Journal of Political Economy 98 No. 5 (October, 1990, part 2): S12650.

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

SYMPOSIUM: WHAT CAN TAX REFORM DELIVER?

King, Robert G., and Sergio Rebelo. Transitional Dynamics and Economic Growth in the Neoclassical Model. American Economic Review 83 No. 4 (September, 1993): 90831. Koester, Reinhard B., and Roger C. Kormendi. Taxation, Aggregate Activity and Economic Growth: Cross Country Evidence on Some Supply-Side Hypotheses. Economic Inquiry 27 No. 3 (July, 1989): 36786. Levine, Ross, and David Renelt. A Sensitivity Analysis of Cross-Country Growth Regressions. American Economic Review 82 No. 4 (September, 1992): 94263. Lindsey, Lawrence. Individual Taxpayer Response to Tax Cuts, 19821984: With Implications for the Revenue Maximizing Tax Rate. Journal of Public Economics 33 No. 2 (July, 1987): 173206. Lindsey, Lawrence. The Growth Experiment. New York: Basic Books, 1990. Lucas, Robert E. Supply-Side Economics: An Analytical Review. Oxford Economic Papers 42 No. 2 (April, 1990): 293316. MaCurdy, Thomas, David Green, and Harry Paarsch. Assessing Empirical Approaches for Analyzing Taxes and Labor Supply. Journal of Human Resources 25 No. 3 (Summer, 1990): 41590. Mankiw, N. Gregory, David Romer, and David N. Weil. A Contribution to the Empirics of Economic Growth. Quarterly Journal of Economics 107 (May, 1992): 40738. Mariger, Randall. Labor Supply and the Tax Reform Act of 1986: Evidence from Panel Data. Federal Reserve Board, Finance and Economics Discussion Series Paper 95-34. Washington, D.C.: Federal Reserve Board, 1995. McKinsey Global Institute; with assistance from Axel Borsch-Supan and our Advisory Committee, Robert Solow, Benjamin Friedman, Zvi Griliches, and Ted Hall. Capital Productivity. Washington, D.C.: McKinsey and Company, 1996. Mendoza, Enrique G., Assaf Razin, and Linda L. Tesar. Effective Tax Rates in Macroeconomics: Cross-Country Estimates of Tax Rates on Factor Incomes and Consumption. Journal of Monetary Economics 34 No. 3 (December, 1994): 297324. Mendoza, Enrique G., Gian Maria MilesiFerretti, and Patrick Asea. On the Ineffectiveness of Tax Policy to Alter Long-Run Growth: Harbergers Superneutrality Conjecture, Federal Reserve Board of Governors. Mimeo, 1996. National Commission on Economic Growth and Tax Reform (NCR). Unleashing Americas Potential. Washington, D.C.: NCR, 1996. 641

Navratil, John. Evidence of Individual Taxpayer Behavior from Panel Tax Return Data. Harvard University. Mimeo, 1994. Nonneman, Walter, and Patrick Vanhoudt. A Further Augmentation of the Solow Model and the Empirics of Economic Growth for OECD Countries. Quarterly Journal of Economics 111 No. 3 (August, 1996): 94353. Pechman, Joseph A. Federal Tax Policy. 5th ed. Washington, D.C.: The Brookings Institution, 1987. Poterba, James. Venture Capital and Capital Gains Taxation. In Tax Policy and the Economy, Vol. 3, edited by Lawrence Summers, 4768. Cambridge, MA: The MIT Press, 1989. Poterba, James M., and Lawrence H. Summers. A CEO Survey of U.S. Companies Time Horizons and Hurdle Rates. Sloan Management Review 37 No. 1 (Fall, 1995): 4353. Ram, Rati. Government Size and Economic Growth: A New Framework and Some Evidence from Cross-Section and Time-Series Data. American Economic Review 76 No. 1 (March, 1986): 191203. Romer, Paul M. Human Capital and Growth: Theory and Evidence. Carnegie-Rochester Conference Series on Public Policy 32 (Spring, 1990): 25186. Romer, Paul M. Increasing Returns and LongRun Growth. Journal of Political Economy 94 No. 5 (October, 1986): 100237. Skinner, Jonathan. Taxation and Output Growth in Africa. Policy, Planning and Research Working Paper 73. Washington, D.C.: The World Bank, August, 1988. Skinner, Jonathan, and Daniel Feenberg. The Effect of the 1986 Tax Reform Act on Personal Saving. In Do Taxes Matter?: The Impact of the Tax Reform Act of 1986, edited by Joel Slemrod, 5079. Cambridge, MA: MIT Press, 1990. Slemrod, Joel. The Economic Impact of the Tax Reform Act of 1986. In Do Taxes Matter?: The Impact of the Tax Reform Act of 1986, edited by Joel Slemrod, 112. Cambridge, MA: MIT Press, 1990. Slemrod, Joel. On the High-Income Laffer Curve. In Tax Progressivity and Income Inequality, edited by Joel Slemrod, 177210. Cambridge: Cambridge University Press, 1994. Slemrod, Joel. What Do Cross-Country Studies Teach About Government Involvement, Prosperity and Economic Growth? Brookings Papers on Economic Activity 2 (1995): 373431. Slemrod, Joel. High-Income Families and the Tax Changes of the 1908s: The Anatomy of

National Tax Journal Vol 49 no. 4 (December 1996) pp. 617-42

NATIONAL TAX JOURNAL VOL. XLIX NO. 4

Behavioral Response. In Empirical Foundations of Household Taxation, edited by Martin Feldstein and James Poterba, 169188. Chicago: University of Chicago Press and NBER, 1996. Solow, Robert M. A Contribution to the Theory of Economic Growth. Quarterly Journal of Economics 70 No. 1 (February, 1956): 6594. Stokey, Nancy L, and Sergio Rebelo. Growth Effects of Flat-Rate Taxes. Journal of Political Economy 103 No. 3 (June, 1995): 51950. Summers, Robert, and Heston, Alan. The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 19501988. Quarterly Journal of Economics 106 No. 2 (May, 1991): 32768.

The World Bank. World Development Report 1988. Oxford: Oxford University Press, 1988. Triest, Robert. The Effect of Income Taxation on Labor Supply in the United States. Journal of Human Resources 25 No. 3 (Summer, 1990): 491516. Triest, Robert. Fundamental Tax Reform and Labor Supply. In Economic Effects of Fundamental Tax Reform, edited by Henry Aaron and William Gale. Washington, D.C.: The Brookings Institution, 1996. Trostel, Philip A. The Effect of Taxation on Human Capital. Journal of Political Economy 101 No. 2 (April, 1993): 32750.

642